Dimensions of Consumer's Perceived Risk in Online Shopping

No.3 YE Naiyi: Dimensions of Consumer’s Perceived Risk in Online Shopping 179age, gender, and education background, experience ofInternet use, experience of online shopping, andintentions to Internet shopping and where they comefrom, to get the demographic information.We take many class times to ask the students toscore the items in the questionnaire. The itemsrepresent the concerns of respondents may have in anonline shopping. In order to keep a generalization, wegave a scenario in which we did not specify the goodpurchased, but we set a scenario of a purchase with thevalue over 100 RMB Yuan in an Internet store in orderto make the online buying action relatively important.After conducting the survey, the data wascollected, coded and examined. We finally obtained336 effective questionnaires as our analyzing cases.Then, we use SPSS 11.5 to conduct exploratory factoranalysis (EFA) to obtain the assumed factor structure,and use LISREL 8.54 to conduct confirmatory factor(CFA) analysis to test the validation of the model.4 Data Analysis and DiscussionsFirstly, we examined the reliability of the wholedata by conducting reliability analysis and obtained theCronbach α of 0.913. Since the α coefficient isrelatively high, then data collected and themeasurement are considered reliable. Than we haveobtained the demographic information of the allrespondent. The results of their demographiccharacteristics are shown in Tab.1.Tab.1Demographic statistics of respondentDemographic statisticsSexMale: 195, 57.7%Female: 143, 42.3%EducationUndergraduate Student: 188, 55.6%Graduate Student: 150, 44.4%AgeRange: 18 ~ 45Average: 24.54ShopperHave experience of shopping online: 57, 16.9%Have no experience of shopping online: 281, 83.1%Attitude to Shopping OnlinePositive: 283, 83.7%Negative: 54, 16.3%After that, we conducted explanatory factoranalysis to find the dimensions of online consumer’sperceived risk. We performed principle componentanalysis and used the orthogonal rotation method ofVarimax with Kaiser Normalization that minimizes thenumber of variables that have high loadings on eachfactor. We obtained the rotated factor loadings as in theTab.2. It presents the factor analysis of the PRD. Thefactor solution explains 63.291% of the variation. Theeigenvalues of all factors exceed 1.0. We have thefollowing explanations about those seven factors.Tab.2 Cronbach α of seven factors from EFAFactors from EFAαFactor 1: e-Store source risk 0.799 0Factor 2: Delivery risk 0.839 9Factor 3: Financial risk 0.752 6Factor 4: Product performance risk 0.680 3Factor 5: Shopping Process risk 0.679 3Factor 6: Privacy risk 0.628 2Factor 7: Asymmetric information risk 0.633 4Factor 1: fraud risk. This factor measures aconsumer’s concerns about seller’s reliability in onlineshopping. Items loading on this factor measure anindividual’s concern to the opportunistic behavior ofthe online retailer. It includes seller’s reliability andpost-services and so on. The highest loading of theitem T1 measures the perception to the web store’sfraud behavior.Factor 2: delivery risk. This factor measures aconsumer’s concern about product delivery. It coversthe concerns about the loss and damage of product, andwrong destination of delivery. The highest loadingmeasures a consumer’s concern about the loss ofonline purchased product.Factor 3: financial risk. This factor measures aconsumer’s concern about monetary loss whenshopping through the Internet. It relates to lowerdiscount in online shopping comparing to traditionalshopping and extra charges of delivery and onlinepayment.Factor 4: process and time loss risk. This factormeasures the easy and convenience of a consumer’sperception about Internet shopping. The highestloading of the item PS1 measures a consumer’sperceived complexity and inconvenience exceeding theexpected process of online shopping.Factor 5: product risk. This factor measures a

180Journal of Electronic Science and Technology of China Vol.2consumer’s concern about the product quality,performance, falseness of a product, and productrelated problem. The highest loading measures aconsumer’s perceived quality.Factor 6: privacy risk. This factor measures aconsumer’s concern about the security of personalinformation. It includes a consumer’s home address,telephone number, e-mail address, and account numberof credit or debit cards. The highest loading of the itemmeasures a consumer’s concern about the misuse ofhome address and telephone number.Factor 7: information risk. This factor measures aconsumer’s perception of asymmetric informationabout both of sellers and products. The highest loadingof the item measures the concern of a consumer aboutthe lack of information about sellers.Tab.3Factor loadings of EFA and coefficients of determination of CFAFactors and Underlying itemsLoadings(EFA)StandardizedCoefficients ofDetermination(CFA)T(CFA)R 2(CFA)Fraud RiskT1: Online information about product is not true. 0.784 0.577 10.813 0.333T2: It is difficult to get support when product fails. 0.716 0.647 12.464 0.419T3: Can’t find the place where to settle disputes. 0.669 0.735 14.737 0.540T4: Web store could disappear after running business in short time. 0.585 0.585 10.987 0.342T5: Fail to keep the promise of post-services. 0.559 0.764 15.526 0.583Delivery RiskD1: The delivered product could be lost. 0.798 0.811 16.861 0.658D2: Delivered the product to a wrong place. 0.774 0.794 16.379 0.631D3: The product is damaged during the delivering. 0.751 0.789 16.229 0.622Financial RiskF1: Traditional stores offer more discount than online store. 0.755 0.450 8.019 0.203F2: Online stores offer discount price but the total cost is not lower. 0.673 0.638 12.081 0.407F3: Online payment will charge extra fees. 0.642 0.780 15.678 0.609F4: Delivering to the home will charge relatively higher fees. 0.572 0.761 15.177 0.580Process and Time Loss RiskPS1: The process of online shopping is complex and inconvenient. 0.788 0.519 8.838 0.296PS2: To deal with PC for accessing Internet will take too much time. 0.687 0.721 12.636 0.514PS3: Information transformation is too slow during online shopping. 0.668 0.710 12.497 0.504Product RiskPT1: The quality of the product is not accepted. 0.797 0.583 10.120 0.340PT2: The product performance is not consistent with the expectation. 0.659 0.620 10.858 0.385PT3: The product may be false and the quality will be poor. 0.567 0.500 8.497 0.250PT4: It is difficult to return when the product is not satisfied. 0.553 0.651 11.473 0.424Privacy RiskPY1: Worrying about that the personal address, telephone numbercould be misused by others.0.727 0.583 10.088 0.340PY2: My e-mail address could be misused by others. 0.718 0.795 13.716 0.632PY3: The account number of my credit or debit card could bemisused by others.Information RiskIN2: The information about online suppliers is not sufficient.0.591 0.454 7.666 0.2070.809 0.715 10.365 0.497IN1: The information about product to be purchased is not sufficient. 0.788 0.658 9.901 0.432

No.3 YE Naiyi: Dimensions of Consumer’s Perceived Risk in Online Shopping 181The presented seven factors structure model hasthe descriptive power about the dimensions ofconsumer’s perceived risk in online shopping. Themodel consists of seven factors extracted from selected24 variables. The Cronbach α coefficient for the 24variables is 0.881 9. That provides the evidence for thereliability of the items measuresEach factor keeps also a relatively higherCronbach α coefficient (for attitude survey) for themodel. Tab.2 lists the Cronbach α coefficients for theseven factor structure model.To examine the validities of the developedconstructs of the model, we also used the results fromthe EFA to conduct confirmatory factor analysis. Weobtained the Coefficients of Confirmation from CFA.The results are shown in the Tab.3.The purpose CFA is to examine the validity of themodel. We used the results of the EFA as ourhypnotized model to confirm the validity of the itemmeasures. We use 24 variables as measures and sevenfactors as latent variables to build a hypothesizedmeasurement model. Then we set the variance of eachlinear equation to 1 to get the complete standardizedestimations of the model. We obtained standardizedcoefficient estimations and R 2 as listed in the Tab.3.The CFA results show that all coefficients aresignificant to represent the linear relationship betweenvariables and the related factors. Some R 2 values areacceptable and some of them are relatively low, thatrepresents that those measures still need to improve.CFA is basically involved specification of ahypothesized model and confirm whether this model isconfirmed by the underlying data.CFA is basically involved specification of ahypothesized model and confirm whether this model isconfirmed by the underlying data. The better themeasures of fit, the more accurate the data in relationto the proposed theory. According to recommendationsof some researchers (Hair et al., 1998), X 2 should notbe significant, X 2 /df is recommend between 1.0~2.0,GFI is better closed to 1, AGFI is recommend morethan 0.8, RMSR is recommend closed to 0; RMSEA isalso recommend closed to 0, NFI is recommend morethan 0.9, CFI is better closed to 1 [12] .Tab.4 lists the results of the goodness of fit fromconfirmatory factor analysis. All indexes of goodnessof fit are acceptable except the X 2 -test. The X 2 -test washighly significant at p=0.00. However, X 2 -tests aresample size dependent and favor complex models thansimple ones. When it is adjusted for degrees offreedom, these and other measures of fit areacceptable.Tab.4 Goodness of fitGoodness of fitX 2 430.601(P = 0.00)X 2 /df 1.864Goodness of Fit Index (GFI) 0.907Adjusted Goodness of Fit Index (AGFI) 0.879Root Mean Square Residual (RMSR) 0.143Root Mean Square Error of Approximation(RMSEA)0.049Comparative Fit Index (CFI) 0.969df: Degrees of Freedom (= 231)5 ImplicationsAlthough the primary purpose of this researchwas to substantiate electronic commerce theory, somemanagerial implications both for e-commerceresearchers and managers can be derived from theresulting research work. Firstly, the research drawsattention to consumer’s perception of risk in Internetshopping. Some researches focused on the consumer’sintentions to buy through retailer’s web site, but failedto identify the consumer’s perceived risk in onlineshopping (George Banabanis and Stefanos Vassileiou,1999) [13] . This research may raise the interests inconsumer’s perceived risk about online shopping andmotivate managers take account into consumer’sconcern when make e-commerce strategies.Perceived risk can be used as overall factor toexplain the risk perception and risk deduction strategyused by consumers. Because outcomes of an exchangeare uncertain, consumers desire to reduce their risk inpurchasing. The dimensions of perceived riskdeveloped in this research are provided for managerswhen they design their e-commerce strategies to satisfythe consumer’s risk reduction need.

182Journal of Electronic Science and Technology of China Vol.2Our research has developed a structure model formeasuring consumer’s perceived risk in onlineshopping. The 24 variables are identified as measuresof consumer’s perceived risk. In the structure model,we find out seven factors as dimensions of theconsumer’s perceived risk in China’s Internet shoppingcontext. The factor one is e-store source risk, factortwo is delivery risk; the factor three is financial risk;the factor four is purchasing process and time loss risk;the factor five is product performance risk; the factorsix is privacy risk; and the factor seven is asymmetricinformation risk.Research draws attention to consumer’sperception of risk in Internet shopping and providesthe specifications about the dimensions of Chineseconsumer’s preserved risk in the online context. It maymotivate managers take account into consumer’sconcern and provide more chances for managers to useour findings for their managerial practice in ane-commerce market environment.AcknowledgementThis paper is supported by the National NaturalScience Foundation of China.References[1] Bauer R A. Consumer Behavior as Risk Taking. In D.F.Cox (ed.), Risk Taking and Information Handling inConsumer Behavior[M]. Cambridge: Harvard UniversityPress, 1960. 389-398[2] Cunningham S M. The Major Dimensions of PerceivedRisk. in Cox, D.F. (Ed.), Risk Taking and InformationHandling in Consumer Behavior[M]. Boston GraduateSchool of Business Administration, Harvard UniversityPress, 1967. 82-108[3] Peter J P, Rayan N J. An investigation of perceived risk atthe brand level[J]. Journal of Marketing Research, 1976,13:184-188[4] Kaplan L B, Jacoby J, Szybillo G. Components ofperceived risk in product purchase: a cross-validation[J].Journal of Applied Psychology, 1974, 59(3): 287-291[5] Fram E H, Grady D B. Internet shoppers: is there a surfergender gap[J]. Direct Marketing, 1997,59 (9): 46-50[6] Jarvenpaa S L, Todd P A. Consumer reactions to electronicshopping on the world wide web[J]. International Journalof Electronic Commerce, 1996, 1(2): 59-88[7] Nyshadham E A. Privacy policies of air travel web sites: asurvey and analysis[J]. Journal Air Transport Management,2000, 6: 143-152[8] McCorkle D E. The role of perceived risk in mail order–catalog shopping[J]. Journal of Direct Marketing, 1990, (4):26-35[9] Engel J F, Blackwell R D. Perceived risk in mail-order andretail store buying[J]. Journal of Marketing Research, 1970,7: 364-369[10] Nena L. Consumers’ perceived risk: sources versusconsequences[J]. Electronic Commerce Research andApplications, 2003, 2: 216-228[11] Sandra M, Shi F B. Consumer patronage and riskperceptions in Internet shopping[J]. Journal of BusinessResearch, 2003, 56(11): 867[12] Hair J F, Anderson R E, Tathan R L, et al. MultivariateData Analysis[M]. EngleWood Cliffs, NJ: Prentice Hall,1998[13] Banabanis G, Vassileiou S. Some attitudinal predictors ofhome shopping through the Internet[J]. Journal ofMarketing Management, 1999, (15): 361-385Brief Introduction to Author(s)YE Naiyi is an associate professor in School of Economicsand Management, Southwest Jiaotong University, China. Hiscurrent research interests include: electronic commerce,marketing, consumer behavior.